Description: In the field of information sciences, attention has been focused on developing mature information retrieval systems that abstract information automatically from the contents of information resources, such as books, images and films. As a subset of information retrieval research, content-based image retrieval systems automatically abstract elementary information from images in terms of colors, shapes, and texture. Color is the most commonly used in similarity measurement for content-based image retrieval systems. Human-computer interface design and image retrieval methods benefit from studies based on the understanding of their potential users. Today's children are exposed to digital technology at a very young age, and they will be the major technology users in five to ten years. This study focuses on children's color perception and color association with a controlled set of digital images. The method of survey research was used to gather data for this exploratory study about children's color association from a children's population, third to sixth graders. An online questionnaire with fifteen images was used to collect quantitative data of children's color selections. Face-to-face interviews investigated the rationale and factors affecting the color choices and children's interpretation of the images. The findings in this study indicate that the color children associated with in the images was the one that took the most space or the biggest part of an image. Another powerful factor in color selection was the vividness or saturation of the color. Colors that stood out the most generally attracted the greatest attention. Preferences of color, character, or subject matter in an image also strongly affected children's color association with images. One of the most unexpected findings was that children would choose a color to replace a color in an image. In general, children saw more things than what were actually represented in the images. However, the children's interpretation ...
Date: August 2003
Creator: Chang, Yun-Ke
Item Type: Thesis or Dissertation
Partner: UNT Libraries